Question

A logistics company wants a lightweight vision app built in Azure AI Foundry. Warehouse workers will take photos of arriving packages on their phones, and the app must automatically read the shipping label text visible in each photo and return it as plain text that can be pasted into a tracking system. The team wants to send the image directly to a deployed model in a single prompt without training a custom model first. Which approach should they use?

Accepted Answer

Deploy a multimodal model and include the package photo as image input in the prompt, asking the model to read and return the label text

Answer

Deploy a text-only generative model and paste a written description of the label into the prompt

Answer

Use Azure Speech in Foundry Tools to transcribe the label from the photo

Answer

Train a custom image classification model to sort packages into categories before extracting text